Variation sets in child directed speech to Argentinian toddlers: effects of SES and type of activity assessed in a naturalistic study

Abstract

A distinct structural feature of Child Directed Speech (CDS) is the use of "variation sets" (Küntay & Slobin, 1996): successive utterances with partial self repetitions. Previous research on mother-child play situations found an impact of socio-economic status (SES) on the quantity and extension of variation sets (Tal & Arnon, 2018). However, every day children are embedded in interactions with multiple people in the context of diverse activities. Hence, we examine the extension and quantity of variation sets in naturalistic at-home multiple participant CDS to Argentinian toddlers. We ask about the effects of SES and the type of the ongoing activity on these characteristics.Participants were 30 socio-economically diverse Argentinian children (8 to 20 months). Families varied regarding mothers? education (primary, secondary, graduate and postgraduate degrees), considered here as a proxy of SES. Children were audio-recorded for 4 hours at-home, without the researchers presence. Transcriptions were done in CHAT format. Each utterance was coded according to the activity and clustered into: 1) structured activities -booksharing, regulated play adult-child conversations-, 2) non-structured activities- feeding, grooming, exploratory object and physical play, household chores, conversations between adults, outings and watching TV. Variation sets were automatically extracted from the CDS provided by all the participants. We conducted linear mixed-effect regression analysis to estimate the effects of mothers education and type of activity on the quantity and extension of variation sets , considering the child as random effect and controlling for age. Results showed an effect of the interaction between mother´s education level and type of activity on the quantity of variation sets: in structured activities families in which mothers? have a post-graduate degree, children heard significantly more variation sets than in families where mothers have a secondary and primary education. Neither predictor showed an effect on the extension of the variation sets.